CONFIDENCE IN MODELING FUTURE CLIMATE - A SOUTHERN-HEMISPHERE PERSPECTIVE

Climate models are essentially surrogates for the real system, in which experiments can be carried out. When these experiments attempt to simulate future climate, the results cannot be compared with the real atmosphere because they involve changes unique in recorded human history. Confidence indicators include model comparison with the real atmosphere for current climate representation and model intercomparison for future climate representation. From a Southern Hemisphere perspective general circulation models (GCMs) reveal some inadequacies in their representation of climate and differ significantly from each other in their response to a CO2 doubling. Representation of drought as a response to sea surface temperature anomaly is shown to be successful but strongly dependent on correct parameterization of land surface exchange processes. More attention to Southern Hemisphere representation is required, particularly because of the likely strong role of the oceans.